Embeddings & Retrieval · 90 minutes
Explain semantic and exact-match retrieval
Explain how embeddings and exact-term matching find candidates for a query, and why the top result is a ranking signal rather than a verified answer.
1.5 h
Used in interview rounds:
Technical Screen (Easy)1
Step 1: The question
Explain how a help center could retrieve D1 for a paraphrased password question and D2 for an exact error code. Describe what embeddings contribute and why the application should not interpret the top result as a verified answer.
Worked search scenario
A small help center has four articles.
| ID | Article |
|---|---|
| D1 | Reset your password using the recovery link |
| D2 | Resolve E104 when a login token expires |
| D3 | Download a billing invoice |
| D4 | Change your profile display name |
Consider three queries against this collection: “I forgot the secret I use to sign in”, “E104”, and “shipping times”. The scenario asks for a reasoned prediction about which retrieval signal would find a useful article. It makes no claim about an actual embedding model's output.
2
Step 2: Hints
Optional. Each hint gives a little more away.
3
Step 3: Your answer
Kept in this browser. Saving your score below keeps a copy with it.
0 / 5000
4
Step 4: Reference answer
5
Step 5: Score yourself
Score each dimension 0, 1, or 2, then save the score to your history. Readiness thresholds are editorial estimates.
Grades your written answer against this rubric with your own AI provider key from Settings. It is a second opinion: the score you save is still yours.
0 of 8 · Scored 0 of 4 dimensions
Ready at 6 of 8, with no 0 in Score interpretation
Kept in this browser until you sign up, then it moves to your account.
Complete this level when you can answer the canonical question in two minutes, rank the three teaching vectors without looking, and explain why the nearest article for a no-answer query is still irrelevant.
Your saved scores
Practice saves and mock interview answers on this question, newest first.
No saved scores yet. Score every dimension, then save.
6
Step 6: Follow-up questions
An interviewer may ask these next.
What is an embedding?
Why might keyword search beat semantic search?
Is a high similarity score confidence?
Does top-5 guarantee five useful results?
Can equal-sized vectors from different models be mixed?
7